water percolating behavior indicated by the water

www.elsevier.com/locate/enggeo

Engineering Geology

Water percolating behavior indicated by the water chemistry within a

decomposed granite slope, central Japan

Masahiro Chigira a,*, Chizuru Imai b, Hideaki Hijikata c

a Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji 611-0011, Japanb Chuo Fukken Consultants, Co., Ltd., 4-11-10 Higashi Nakashima, Higashi Yodogawa, Osaka 533-0033, Japan

c Nippon Tochi-Tatemono Co. Ltd., 1-4-4 Kasumigaseki, Chiyoda-ku, Tokyo 100-0013, Japan

Received 15 April 2005; received in revised form 14 December 2005; accepted 15 December 2005

Available online 3 February 2006

Abstract

Monitoring rainfall, discharge amount, and the chemistry of water from just beneath the groundwater table yielded a conceptual

model of water percolation behavior within a decomposed granite slope in central Japan, providing a clue to evaluate a landslide

hazard potential by rainstorms. When rainfall is strong enough, more than about 50 mm a day in this case, air pressure in the vadose

zone probably increases temporarily because of the downward movement of a wetting front, pushing down the water of the

capillary fringe. The capillary water has SiO2 concentrations controlled by the precipitation of cristobalite. The water is rich in Ca,

which is supplied from plagioclase with SiO2. After the pushing down of the capillary water, rainwater with soil water in the

unsaturated zone reaches the groundwater surface. The cation concentrations in the unsaturated soil water and the capillary water

are affected by ion exchange.

D 2006 Elsevier B.V. All rights reserved.

Keywords: Weathering; Water filtration; Capillary water; Rainstorm

1. Introduction

Rainwater percolation within weathering profiles is

very important in various scientific phenomena and en-

gineering, such as weathering, geochemical erosion,

hydrological budget within a catchment, pollutant mi-

gration, landslide generation, and so on, but has not yet

been sufficiently clarified. This is partly because weath-

ering profiles are specific to rock types and weathering

conditions, and their characteristics and formative pro-

cesses have been studied only recently and have not yet

produced definitive results. Water filtration through va-

dose zone has been studied in the fields of soils science

0013-7952/$ - see front matter D 2006 Elsevier B.V. All rights reserved.

doi:10.1016/j.enggeo.2005.12.002

* Corresponding author. Fax: +81 774 38 4100.

E-mail address: [email protected] (M. Chigira).

and hydrology, and downward advancement of a wetting

front (Bodman and Colman, 1943; Rubin and Steinhardt,

1964), its destabilization (Hill and Parlange, 1972;

Raats, 1973; Philip, 1975; Hillel and Baker, 1988), and

air entrapment beneath the wetting front are suggested

(Bianchi and Haskell, 1966; Dixon and Linden, 1972;

Kayane and Kaihotsu, 1988). The role of macropore for

quick flow is reviewed by Beven and Germann (1982).

However, direct response of the groundwater to rainfall

event has not reported except for groundwater level

(Bianchi and Haskell, 1966; Linden and Dixon, 1973;

Kayane and Kaihotsu, 1988).

Granite, which is the rock type for this study, is

commonly weathered as deep as tens of meters (Chi-

gira, 2001), and has been prone to shallow landslides

by rainstorms all over the world (Durgin, 1977; Calca-

84 (2006) 84–97

M. Chigira et al. / Engineering Geology 84 (2006) 84–97 85

terra et al., 1996). Therefore, water filtration behavior

within a weathering zone of granite is essential for the

stability evaluation and prediction of rain-induced shal-

low landslide. However, very little is known on this

issue. Weathering of granitic rocks has been studied

from engineering geological view points (Moye, 1955;

Ruxton and Berry, 1957; Dearman, 1974), geochemical

or petrological view point (Wilson, 1966; Kimiya,

1981; Dong et al., 1998; Fritz, 1988; Murphy et al.,

1998; White et al., 1998, 1999). However, few studies

were made on the water filtration through weathered

granite (Asano et al., 1996, 2003, 2004; Shimada et al.,

1992, 1993). Recently, the Japanese Meteorological

Agency developed a sophisticated alarming system

against the hazard of landslide and debris flow by

using a soil moisture index, which is based on the

amount of soil moisture within three conceptual tanks

connected in series (Okada et al., 2000). The tanks

accumulate rainwater, which is calculated hourly on

the basis of AMEDAS (Automated Meteorological

Data Acquisition System) and radar data. The three

tanks simulate the surface shallow layer, the intermedi-

ate layer, and the deeper layer, respectively, but they are

merely conceptual. In order to predict the timing of

landslides, we need to clarify the infiltration behavior of

rainwater through weathering zones.

This study intends to clarify the downward perco-

lating behavior of rainwater and soil water through the

vadose zone to the groundwater table by investigation

of the chemistry of the discharge through drill holes

along the groundwater table within a decomposed gran-

ite slope.

2. Study site

The investigation site is located in the southern part of

Shiga Prefecture, central Japan. This site is a hilly area

with elevations from 450 to 650 m above sea level, and is

underlaid by Cretaceous Shigaraki granite (Collabora-

tive Research Group for the Granites around Lake Biwa,

1982). The landform of the site has low relief with a

relative height of less than 100 m from flat river bottoms.

The granite is weathered as deep as more than 10 m, and

almost all outcrops above the river bottoms consist of

decomposed granite except for sporadically exposed

core stones, while beneath stream bottoms less weath-

ered and sound granite rock is expected. This area was hit

by a heavy rainstorm in 1953 and devastating, widely

distributed landslides were induced.

The study slope, which is at an elevation of about

450 m, was an artificial slope excavated presumably 8 m

(detailed data are not available) from 1987 to 1988, 15

years before the start of the monitoring (Fig. 1). Four

years after the excavation, low-angle drainage drill

holes were excavated from the bottom of the cut

slope; the drill holes are 20 m long, and set 5 m

apart. We monitored discharging water from these

drill holes (numbers #1 to #4). The cut slope is 27 m

high with inclinations of 458 to 258; the lower part is

steeper than the upper. The deepest end of these drill

holes was about 10 m below the slope surface in a

vertical distance. PVC casings, 8.5 cm in outer diameter

and with openings in the interval deeper than 4 m, were

inserted into these holes. These drill holes have yielded

water up to 1.6 L/min, and only the #3 hole sometimes

dried up. We tried to extract water from different inter-

vals from the #1 hole by separating the length of the

hole with rubber packers and setting intakes along the

axis of the drill hole, but after the packers had been set

the discharge stopped, which indicates that the ground-

water table is along the bottom of the drill hole and

never reached the drill hole’s axis. The groundwater

table was also observed within a vertical drill hole in

the middle of the slope (Fig. 1).

Only grasses grow on the lower part and coniferous

trees planted just after the excavation have grown on

the upper part of the study slope. The slope is underlaid

by decomposed granite, which can be seen in patches.

The granite is coarse-grained and consists mainly of

quartz, orthoclase, plagioclase, and biotite. The surface

was loosened to a depth of 0.5 to 1 m, which was

checked by portable cone penetration tests.

We measured temperature within the drill holes and

obtained temperatures of 13.7 to 14.2 8C in the depth of

the drill holes while the outside temperature was 18 8C.The temperatures deeper than 7 m were almost con-

stant, with minimum values at 13, 12, and 16 m for #2,

#3, and #4 holes, respectively. These findings indicate

that at the temperature-minimum points, water presum-

ably comes from deeper than 7 m.

Mean annual temperature at the site is 12.1 8C, theaverage in winter is 2 8C, the average in summer is

22.4 8C, and the mean annual precipitation is about

1500 mm (Hikone meteorological station of the Japan

Metrological Agency).

3. Methods

The monitoring interval was April 3 to December 1,

2003, during which the precipitation and discharge

were measured and water was sampled from the drill

holes at intervals from 1 to 7 days. The precipitation

was monitored by using a tipping bucket rain gauge

with a volume of 0.5 mm set on the middle of the slope.

Fig. 1. Topographic map and a cross section of the monitored slope. Dashed lines indicate low-angle drill holes, in which #1 to #5 are numbered. A

low-angle solid line in the cross section indicates a drill hole #1.

M. Chigira et al. / Engineering Geology 84 (2006) 84–9786

The rain gauge could not be used from August 30 to

September 28, and during this interval, the precipitation

data obtained at a Ministry of Land, Infrastructure, and

Transportation station 1.4 km to the southeast of the

study site were used. The drill holes for the monitoring

were #2, 4, and 5. The discharge was measured man-

ually at the time of water sampling. pH and electrical

conductivity were measured at the site by using a glass

electrode (IM-22P equipped with a glass electrode of

GST-2729C made by DKK-TOA Corporation) and

Twin Cond B-173 made by Horiba, Ltd. The accuracy

of the pH meter was 0.1 and the accuracy of the

conductivity meter was 4 AS/cm. Water samples were

filtered through a membrane with 0.45 Am-pore size

and were analyzed in the laboratory by using the mo-

lybdenum blue method for SiO2 and ion chromatogra-

phy for Na, K, Ca, Mg, Cl, and SO4. The colorimetric

spectrometer and ion chromatography used were HIC-

6A and UV-2200A of Shimadzu Corporation, respec-

tively. The detection limit of the ion chromatography

was less than 10 ppb, and the experimental errors of the

molybdenum blue method were less than 1%. Rainwa-

ter was also sampled and analyzed for these constitu-

ents: it was collected in a polypropylene bottle through

a funnel with a ping-pong ball; the ball prevented the

water from evaporation. Rainwater accumulated in a

few to 7 days was sampled and analyzed 3 to 5 times

per month from June to November.

4. Results and discussion

4.1. Discharge

The rainfall amount from April 1 to December 1

was 569 mm, with a maximum daily precipitation of

80.5 mm on August 14. The discharge amounts of

Fig. 2. Relationships among the discharge amounts from drill holes

#2, #4, and #5. The discharges from #2 and #5 are positively related

to the discharge fron #4.

Fig. 3. Relationships between antecedent rainfall amounts and the

discharge amounts from drill holes #2, #4, and #5. Antecedent periods

are 10, 20, and 30 days from the top to the bottom, respectively.


three drill holes were different probably because of the

detailed hydrogeological structure near each drill hole,

but they were positively related to each other (Fig. 2),

indicating that the recharge response to rainfall events

had a similar pattern among these three drill holes,

but, in more detail, the timing of the discharge max-

imum after rainfall events varied from 1 to 7 days

after rainfall events, as will be discussed later. The

discharges from drill hole #4 were always the largest,

and ranged 280–1500 mL/min; those from drill hole

#2, from 50–1200 mL/min; and those from drill hole

#5, 60–500 mL/min. The variation in discharge

amount is influenced by antecedent rainfall, the effect

of which is shown in Fig. 3 as diagrams of discharge

versus antecedent rainfall over 10, 20, and 30 days.

The correlation coefficient is largest at 20 days for #2

(0.82) and 30 days for #4 (0.55) and #5 (0.6) (Table

1), suggesting that rainfall effects on water discharge

remain for tens of days. Fig. 3 also shows that as

cumulative rainfall within the above time intervals

approach zero, the discharges from #2 and #5 con-

verge to zero while the discharge from #4 does not,

which suggests that the relative amount of stationary

lateral groundwater to the amount of recharge water

from soil water is larger for #4 than for #2 or 5.

4.2. Chemistry

Averages and standard deviations of each element

are shown in Table 2 for each drill hole and for accu-

mulated rainwater.

Comparison of the chemistry of rainwater and that of

discharge water showed that rainwater was richer than

discharge water in potassium, indicating that potassium

is adsorbed to weathered granite or consumed by bio-

genic activity during percolation. Chlorine was present

in discharge at average concentrations of about the

average plus one standard deviation of Cl in rainwater,

and its standard deviations were much smaller than

those in discharge water. This finding suggests that the

Cl variations in discharge water reflect the variations in

rainwater, which is smoothed during percolation. More

sulfate was contained in discharge water than in rain-

water, though the explanation for this has not been

clarified yet. SiO2 was not detected in rainwater.

Table 1

Correlation coefficients between the antecedent rainfalls and dis-

charges

Drill hole Interval (days)

10 20 30

#2 0.72 0.82 0.68

#4 0.27 0.46 0.55

#5 0.32 0.45 0.60


The interrelationships among the discharges for these

respective elements are shown in Fig. 4. All components

except Mg showed good positive relationships among

the three drill holes, indicating that they behaved in a

similar way in these drill holes. Electrical conductivities

ranged from 35 to 90 AS/cm, and showed positive

relationships among the three drill holes. PH values

ranged from 6.2 to 7.4 and were lower in the water

from drill hole #4 than in that from other drill holes.

4.2.1. SiO2

Silica concentrations ranged from 0.33 to 0.52mmol/L

with fluctuations: gradual increases followed by abrupt

decrease (Fig. 5). This fluctuation interval appears to

have a lower limit, which approximately coincides with

the solubility of cristobalite, 0.34 mmol/L, calculated

from the equation of Fournier (1985). Dove (1995)

reported solubilities of 0.23 mmol/L for cristobalite

and 0.30 mmol/L for tridymite at 258, but the solubilitiesof these facies are not well established. SiO2 mainly

comes from plagioclase of the decomposed granite,

and Na and Ca are both released with it, but the total

concentrations of Na and Ca are not related to the con-

centrations of SiO2 (Fig. 6), which also indicates that

some reactions occur after the dissolution of plagioclase.

Shimada et al. (1993) monitored silica concentration

once a day in shallow groundwater at a weathered

granite area near our study site for more than one

year, and reported that the silica concentrations were

Table 2

Chemistry of discharge water from the drill holes and rainwater (mmol/L e

Na K Mg Ca Cl

Hole 2 Ave 0.267 0.001 0.007 0.051 0.035

Sdv 0.016 0.003 0.007 0.015 0.004

4 Ave 0.358 0.002 0.019 0.089 0.044

Sdv 0.025 0.002 0.020 0.012 0.003

5 Ave 0.300 0.002 0.010 0.111 0.038

Sdv 0.017 0.002 0.002 0.028 0.003

Rain Ave 0.014 0.017 0.002 0.011 0.018

Sdv 0.015 0.024 0.002 0.010 0.017

Monitoring interval: April 3 to December 1, 2003. We chose SiO2, Ca, Na

weathering profile: SiO2 is not contained in rainwater, and Ca, Na, and Mg

the rainwater. Ave: average; sdv: standard deviation.

nearly constant at 0.3 to 0.33 mmol/L with temporary

dilutions caused by heavy rainstorms. Although they

did not provide an explanation for the constant values,

it is strongly suggested that SiO2 concentration is con-

trolled by the solubility of cristobalite. However, the

constant values of SiO2 found by Shimada et al. (1992,

1993) were upper limits instead of lower limits, which

was the case in our study. This is probably due to the

fact that our study slope is an artificial cut slope, where

reactive decomposed granite is exposed to filtrating

water, whereas the study slope of Shimada et al.

(1992, 1993) was a natural slope, where mineral sur-

faces have been exposed to filtrating water for long

time. Plagioclase with a fresh reactive surface is dis-

solved to release SiO2 into water, which is easily su-

persaturated with quartz, cristobalite, and often

amorphous silica (Busenberg and Clemency, 1976),

which suggests that once water is supersaturated with

cristobalite, it approaches its equilibrium. On the other

hand, plagioclase is expected to be dissolved more

slowly under the condition of a natural slope; the

water is not supersaturated with cristobalite. Chigira

et al. (2002) studied the weathering profiles of rhyolitic

vapor-phase crystallized ignimbrite and found that tri-

dymite, the major constituent of the ignimbrite, is dis-

solved and cristobalite is formed at the weathering

front. This finding supports the idea that SiO2 concen-

trations in shallow groundwater in acidic igneous rocks

are controlled by the solubility of cristobalite.

4.2.2. Ca, Na and Mg

Calcium concentrations ranged from 0.02 to 0.22

mmol/L, with the highest in #5 and the lowest in #2.

Ca showed temporal changes, responding to rainfall

events; these will be discussed later. Asano et al.

(1996) analyzed the groundwater in granite slopes near

our study site and reported Ca concentrations less than

0.03 mmol/L, which almost coincided with the through-

xcept for pH and EC (AS/cm))

SO4 SiO2 pH EC (AS/cm) Q (mL/min)

0.022 0.389 6.8 47.9 255

0.002 0.041 0.1 4.5 170

0.024 0.384 6.4 66.4 721

0.002 0.038 0.1 4.4 264

0.026 0.419 6.8 66.2 177

0.003 0.048 0.2 6.6 77

0.008 Nd 4.3 40.1 –

0.009 Nd 0.4 30.4 –

, Mg, and EC for examining the water filtration behavior within the

were contained in higher concentrations in the discharge water than in

Fig. 4. Interrelationships among three discharges for SiO2, Ca, Na, and Mg concentrations; electrical conductivity; and pH.


fall water in the forest, which was 0.005–0.03 mmol/L.

In our study site, on the other hand, Ca in the rainfall

averaged 0.01+0.01 mmol/L, and Ca in the discharge

water averaged 0.086 mmol/L with a minimum of

0.02 mmol/L, indicating that a significant amount of

Ca came from the decomposed granite. This may be

due to the fact that the study slope was an artificial cut

slope, exposing reactive decomposed granite. Busenberg

Fig. 5. Temporal changes in SiO2 concentration. Cristobalite solubility is shown as a line of 0.34 mmol/L.


and Clemency (1976) showed that Ca concentration

easily exceeds 0.4 mmol/L through batch dissolution

experiments with plagioclase.

Sodium concentrations ranged from 0.15 to 0.45 mmol/L,

with the maximum at #4 and the minimum at #2. The

temporal changes of Na were not related to rainfall

events. Paces (1972) obtained 0.1–0.4 mmol/L of Na

from groundwater in a Bohemian granite area, and

Asano et al. (1996, 2004) reported 0.1–0.4 mmol/L of

Na in shallow groundwater in granite. The Na concen-

trations in our study site are consistent with these data.

However, Ca concentrations were much higher in our

study site than those reported for other natural slopes,

which may be attributed to fact that the release rate of Ca

from plagioclase is much higher than that of Na (Lasaga,

1998) and to the ion exchange reactions discussed later.

Magnesium concentrations ranged from 0.004 –

0.023 mmol/L, with the maximum at #4 and the mini-

mum at #2. Asano et al. (1996, 2004) reported low Mg

concentrations in the same order from groundwater in the

granite area at the Tanakami Mountains near this site.

4.3. Ion exchange reactions suggested by the relations

among cation concentrations

Cations Na, Ca, and Mg in the discharge water

showed clear interrelationships, indicating that cation

exchange reactions occurred during the filtration of soil

water (Fig. 6). In the Na–Ca diagram, water from #4 is

plotted along a line with an inclination of 2 :1, and the

water from #5 and #2 is plotted along a line with an

almost constant Na concentration. In the Na–Mg dia-

gram, water from #4 is plotted along a line with an

inclination of 2 :1, and the water from #5 and #2 is

plotted along a line with a constant Na concentration.

Ion exchange equilibrium between monovalent cation

and divalent cation produces a plot of these concentra-

tions in water on a line with an inclination of 2 :1

(Drever, 1997), so the above results indicate that ion

exchange equilibrium is established for Na–Mg and

Na–Ca for the water from #4 but not for the water

from #5 and #2. The reason for this discrepancy is

not clear, but it may be related to the fact that the rate

of stationary lateral groundwater flow to the recharge

water from soil water is larger for #4 than for #2 and #5,

as mentioned earlier. Only downward filtration through

a vadose zone might not be enough for the establish-

ment of ion exchange equilibrium for Na–Ca or Na–

Mg. In a diagram of Ca–Mg, the water from #4 and the

water from #5 and #2 are plotted along different two

lines, but these two lines have an inclination of 1 :1.

This indicates that ion exchange equilibrium is estab-

lished for Ca–Mg for the water from all the drill holes.

4.4. The change in water chemistry and discharge

amounts responding to rainfall events

During the monitoring interval, SiO2 fluctuated;

showing slow increase and abrupt decrease, and Ca

also fluctuated but with apparently reverse patterns.

The fluctuation patterns are shown for SiO2 and Ca

along with discharge amounts in Fig. 7. The timing of

the abrupt decrease of SiO2 coincided with the heavy

rainfall events, though not all the heavy rainfall events

Fig. 6. Interrelationships among cations and SiO2.


were detected to be accompanied by an abrupt decrease

of SiO2: for example, abrupt decrease occurred at the

events of June 24–25 (101.5 mm), July 13 (46 mm),

September 24 (60 mm), and October 13–14 (54.5 mm),

as indicated by the arrows in Fig. 7. SiO2 decrease

occurred within 1 to 2 days after rainfall events, and

discharge increased contemporaneously and reached

the maximum after the minimum of SiO2 concentra-

tions had been reached. Ca concentrations also

changed, though less remarkably than SiO2 concentra-

tions, in response to rainfall events. At the above

rainfall events, Ca increased temporarily in the dis-

charge from #2 and #5, whereas the change in the

water from #4 was not significant. We could not rec-

ognize the correspondence of precipitation, SiO2

change, and Ca change precisely except for the obvious

ones as above. It is probably due to the fact that

rainwater filtrates and replaces the water in the vadose

zone sequentially as is discussed below. So, when

rainfall occurs one after another like August, the corre-

spondence apparently disappears.

Water sampling intervals in our study were from 1 to

4 days, and not all data can be used to trace the detailed

change in water chemistry according to the rainfall

events; the most dense sampling before, during, and

after rainfall events was made at a rainfall event in June,

51 mm from 8 pm on 23 to 8 am on 24, and 51.5 mm

from 9 pm on 24 to 5 am on 25 (Table 3). Sampling

dates before, during, and after this event were 9 am on

20, 9 am on 23, 9 am on 24, 3 pm on 24, 10 am on 25,

10 am on 26, and 10 am on 27 (Table 3). The results are

shown in Fig. 8, in which hourly precipitation, dis-

charge, SiO2 concentration, Ca concentration, and elec-

trical conductivity are shown. The concentrations are

Fig. 7. Fluctuation patterns of SiO2, Ca, and discharges. Arrows indicate abrupt decreases in SiO2 concentration.


shown as the difference from the average of each

composition (Table 2). The changes in these parameters

were conspicuous in the discharges from #5 and #2 but

less significant in #4. SiO2 decreased by 0.04 to

0.06 mmol/L at the event and recovered to pre-event

values on 27, 2 days after the event, for all the drill

holes. Ca increased by 0.04 to 0.08 mmol/L at the event

and recovered to the pre-event values on 27 for the

Table 3

Dense sampling time in June and July

20-Jun 23-Jun 24-Jun 25-Jun 26-Jun 27-Jun 30-Jun 3-Jul 4-Jul 5-Jul 7-Jul

9:00 9:00 9:00 15:00 10:00 10:00 10:00 9:00 9:00 9:00 11:00 8:00

The data are shown in Fig. 7.


water of #5 and #2, at which time no Ca change in the

water from #4 was observed. Electrical conductivity

apparently responded to the rainfall event, but the

changes were almost within the range of accuracy.

The increase in discharge from these three drill holes

started at the time of the rainfall event and continued to

June 30, much farther than the maximum or minimum

of the chemical components and electrical conductivity.

The changes in SiO2 and Ca concentrations, electri-

cal conductivities, and discharges detected as outlined

above were from intermittent sampling of a 6-h interval

Fig. 8. The response of SiO2, Ca, electrical cond

at the minimum, so their time resolution was not suf-

ficient for detailed analysis. However, continuous mon-

itoring of the electrical conductivity and discharge

performed for the same drill holes in 2004 indicates

that the changes in these parameters very clearly cor-

respond to rainfall events. The electrical conductivity

meter used at this time was WM-22EP (accuracy

b1 AS/cm; DKK-TOA Corporation). An example of

the continuous monitoring is shown for #5 in Fig. 9. In

the case shown in Fig. 9, the following changes occurred

sequentially in response to a rainfall event of 47.5 mm

uctivity, and discharge to a rainfall event.

Fig. 9. Continuous monitoring results on discharge and electrical conductivity for the discharge from drill hole #5 in June, 2004.


precipitation. First, EC dropped and discharge increased

temporarily within 1 h after the peak of rainfall, then EC

gradually increased and reached its peak 20 h after the

rainfall peak, and then decreased. In contrast, discharge

increased further and reached its maximum 50 h after the

rainfall peak. Similar response in EC and discharge was

observed for almost all the rainfall events with cumula-

tive amounts more than 30 mm. The water sampling for

the chemical analysis in this study probably did not catch

the first diluted water. However, the consecutive increase

and decrease of EC were detected in accordance with the

changes in SiO2 and Ca concentrations. These findings

strongly indicate that low-SiO2, high-Ca, and high-EC

water is supplied to the groundwater at strong rainfall

events.

4.5. Water percolation behavior within the vadose zone

indicated by the response of water chemistry to rainfall

events

The changes in water chemistry described above

were detected from the water that emerges just beneath

the groundwater table; hence they can be explained by

the water percolation behavior from the ground surface

to the groundwater table through the vadose zone (Fig.

10). First, the spontaneous drop of EC and the incre-

ment in discharge amount at a rainfall event can be

explained as quick inflow of rainwater to the ground-

water table through fast paths like cracks (B in Fig. 10).

This water disappears immediately after the end of a

rainfall event. The next water, which has low SiO2 and

high Ca, cannot be explained by the mixture of con-

centrated soil water and diluted rainwater because Ca

concentration is high at first. It must be concentrated

soil water pushed down by the air pressure in the

vadose zone, which is induced by the descending wet-

ting front as was suggested by Bianchi and Haskell

(1966) (C in Fig. 10). The air pressure propagation

would not occur if fast paths like cracks were connected

to the vadose zone beneath the wetting front, because

the pressure would be vented (Dixon and Linden,

1972). Therefore, the fast flow mentioned above

would be due to that cracks crossing the drill hole in

its shallower part. Because the water we monitored was

from just beneath the groundwater table, that water that

is pushed down is from the capillary fringe above the

Fig. 10. Diagram explaining the changes of discharge, EC, and Ca and SiO2 concentrations (lower) by the percolation of soil water and rainwater

(upper). The upper left box schematically shows the average concentrations of SiO2 and Ca in soil water.


groundwater table. The water from this capillary fringe

has low SiO2 concentrations close to equilibrium with

cristobalite; thus, SiO2 concentrations in the capillary

fringe may be controlled by the precipitation of cristo-

balite. After this pushing down of the capillary fringe

water, high SiO2, low Ca water, which seems to be soil

water supplied with SiO2 and Ca from plagioclase in

the decomposed granite, follows. With time from rain-

fall events, rainwater itself goes down to the ground-

water table, probably mixing with soil water and

decreasing EC and increasing discharge (D in Fig.

10). Subsequently, discharge again decreases.

Yoshimura et al. (1992) monitored electrical conduc-

tivity in the discharge water from a cavern in limestone,

and found that just after strong rainfalls, electrical

conductivity increases before the discharge peak.

They attributed the first concentrated discharge water

to the soil water in the vadose zone, which is pushed

down by the infiltrated rainwater.

5. Conclusions

We monitored rainfall, discharge amount, and the

chemistry of water from sub-horizontal drill holes just

beneath the groundwater table within a decomposed

granite slope for 10 months in central Japan. When

rainfall was sufficiently strong, more than about 30 mm

per day, characteristic water came out first, then more

dilute but greater water flux came out, and then flux

decreased again. The first characteristic water appears

to be the capillary water pushed down by the air-pressure

propagation induced by the downward movement of a

wetting front. The SiO2 concentration of this water was

apparently controlled by the solubility of cristobalite and


was rich in Ca, which is supplied from plagioclase with

SiO2. After the pushing down of the capillary water,

rainwater with soil water in the unsaturated zone reaches

the groundwater table, which was indicated by further

increase in the water flux and decrease in the electrical

conductivity in the discharges from the drill holes. The

cation concentrations in the unsaturated soil water and

the capillary water are affected by ion exchange.

Acknowledgments

We are grateful to H. Suwa and R. Sidle of the

Disaster Prevention Research Institute, Kyoto Universi-

ty for their discussion. K. Nishiyama of Tokushima

University and S. Nakahara helped us to make the

water sampling system. T. Saito of the Disaster Preven-

tion Research Institute, Kyoto University helped to do

field experiments. We thank all of them. We used Sci-

ence Research Fund from the Ministry of Education,

Culture, Sports, Science and Technology (Grant number

15340169).

References

Asano, Y., Ohte, N., Kobashi, S., 1996. Changes of hydrochemical

processes along forest succession—catchment observation on the

three different successional stages. Bulletin of the Kyoto Univer-

sity Forests (68), 25–42.

Asano, Y., Uchida, T., Ohte, N., 2003. Hydrologic and geochemical

influences on the dissolved silica concentration in natural water in a

steep headwater catchment. Geochimica et Cosmochimica Acta 67.

Asano, Y., Ohte, N., Uchida, T., 2004. Sources of weathering-derived

solutes in two granitic catchments with contrasting forest growth.

Hydrological Processes 18, 651–666.

Beven, K., Germann, P., 1982. Macropores and water flow in soils.

Water Resources Research 18, 1311–1325.

Bianchi, W.C., Haskell, E.E.J., 1966. Air in the vadose zone as it

affects water movements beneath. Water Resources Research 2,

315–322.

Bodman, G.B., Colman, E.A., 1943. Moisture and energy conditions

during downward entry of water into soils. Soil Science Society of

America Proceedings, vol. 8, pp. 116–122.

Busenberg, E., Clemency, C.V., 1976. The dissolution kinetics of

feldspars at 25 8C and 1 atm CO2 partial pressure. Geochimica

et Cosmochimica Acta 40, 41–49.

Calcaterra, D., Parise, M., Dattola, L., 1996. Debris flows in deeply

weathered granitoids (Serre Massif-Calabria, Southern Italy). In:

Senneset, K. (Ed.), Proceedings, 7th International Symposium on

Landslides. Balkema, Trondheim, pp. 171–176.

Chigira, M., 2001. Micro-sheeting of granite and its relationship

with landsliding specifically after the heavy rainstorm in June

1999, Hiroshima Prefecture, Japan. Engineering Geology 59,

219–231.

Chigira, M., Nakamoto, M., Nakata, E., 2002. Weathering mechan-

isms and their effects on the landsliding of ignimbrite subject to

vapor-phase crystallization in the Shirakawa pyroclastic flow,

northern Japan. Engineering Geology 66, 111–125.

Collaborative Research Group for the Granites around Lake Biwa,

1982. Granitic masses around Lake Biwa, Southwest Japan—on

the granites in the Koga area. Journal of the Geological Society of

Japan 88, 289–298.

Dearman, W.R., 1974. Weathering classification in the characterisa-

tion of rock for engineering purposes in British practice. Bul-

letin of the International Association of Engineering Geology 9,

33–42.

Dixon, R.M., Linden, D.R., 1972. Soil air pressure and water infil-

tration under border irrigation. Soil Science Society of America

Proceedings, vol. 36, pp. 948–953.

Dong, H., Peacor, D.R., Murphy, S.F., 1998. TEM study if progres-

sive alteration of igneous biotite to kaolinite throughout a weath-

ered soil profile. Geochimica et Cosmochimica Acta 62–11,

1881–1887.

Dove, P.M. (Ed.), 1995. Kinetic and Thermodynamic Controls on

Silica Reactivity in Weathering Environments, Reviews in Min-

eralogy, vol. 31. Mineralogical Society of America, Washington,

D.C. 235–290 pp.

Drever, J.I., 1997. The Geochemistry of Natural Waters: Surface and

Groundwater Environments, 3rd ed. Prentice-Hall, New Jersey.

436 pp.

Durgin, P.B., 1977. Landslides and the weathering of granitic rocks.

Geological Society of America Reviews in Engineering Geology

3, 127–131.

Fournier, R.O., 1985. The behavior of silica in hydrothermal solu-

tions. In: Berger, B.R., Bethke, P.M. (Eds.), Geology and Geo-

chemistry of Epithermal Systems. Reviews in Economic Geology.

Society of Economic Geologists, pp. 45–61.

Fritz, S.J., 1988. A comparative study of gabbro and granite weath-

ering. Chemical Geology 68, 275–290.

Hill, D.E., Parlange, J.Y., 1972. Wetting front instability in layered

soils. Soil Science Society of America Journal 36, 697–702.

Hillel, D., Baker, R.S., 1988. A descriptive theory of fingering during

infiltration into layered soils. Soil Science 146, 51–56.

Kayane, I., Kaihotsu, I., 1988. Some experimental results concerning

rapid water table response to surface phenomena. Journal of

Hydrology 102, 215–234.

Kimiya, K., 1981. Distribution of weathering crusts and their forma-

tive period in the Mikawa Highland, Central Japan. Journal of the

Geological Society of Japan 87, 91–102.

Linden, D.R., Dixon, R.M., 1973. Infiltration and water table effects

of soil air pressure under border irrigation. Soil Science Society of

America Proceedings, vol. 37, pp. 94–98.

Lasaga, A.C., 1998. Kinetic Theory in the Earth Sciences. Princeton

University Press, Princeton.

Moye, D.G., 1955. Engineering geology for the Snowy Mountain

scheme. Journal of Institution of Enginers, Australia 27, 287–298.

Murphy, S.F., Brantley, S.L., Blum, A.E., White, A., Dong, H., 1998.

Chemical weathering in a tropical watershed, luquillo mountains,

Puerto Rico: II. Rate and mechanism of biotite weathering. Geo-

chimica et Cosmochimica Acta 62, 227–243.

Okada, K.,Makihara, Y., Shimpo,A., Nagata, K., Kunitsugu,M., Saito,

K., 2000. Soil water index (in Japanese). Tenki 48, 349–356.

Paces, T., 1972. Chemical characteristics and equilibration in natural

water-felsic rock–CO2 system. Geochimica et Cosmochimica

Acta 38, 217–240.

Philip, J.R., 1975. Stability analysis of infiltration. Soil Science

Society of America Proceedings, vol. 39, pp. 1042–1049.

Raats, P.A.C., 1973. Unstable wetting fronts in uniform and nonuni-

form soils. Soil Science Society of America Proceedings, vol. 37,

pp. 681–685.


Rubin, J., Steinhardt, R., 1964. Soil water relations during rain

infiltration: III. Water uptake at incipient ponding. Soil Science

Society of America Proceedings, vol. 28, pp. 614–619.

Ruxton, B.P., Berry, L., 1957. Weathering of granite and associated

erosional features in Hong Kong. Bulletin of the Geological

Society of America 68, 1263–1292.

Shimada, S., Ohte, N., Tokuchi, N., Suzuki, M., 1992. SiO2 concen-

tration of groundwater and streamwater in a small forested water-

shed. Journal of the Japan Society of Hydrology and Water

Resources 5 (2), 3–11.

Shimada, Y., Ohte, N., Tokuchi, N., Suzuki, M., 1993. A dissolved

silica budget for a temperate forested basin. In: Peters, N.E.,

Hoehn, E., Leibundgut, C., Tase, N., Walling, D.E. (Eds.), Tracers

in Hydrology. International Association of Meteorology and At-

mospheric Sciences, pp. 79–88.

White, A.F., Blum, A.E., Schulz, M.S., Vivit, D.V., Stonestrom, D.A.,

Larsen, U.S., Murphy, S.F., Eberl, D., 1998. Chemical weathering

in a tropical watershed, luquillo mountains, Puerto Rico: I. long-

term versus short-term weathering fluxes. Geochimica et Cosmo-

chimica Acta 62, 209–226.

White, A.F., Bullen, T.D., Vivit, D.V., Schulz, M.S., Clow, D.W.,

1999. The role of disseminated calcite in the chemical weather-

ing of granitoid rocks. Geochimica et Cosmochimica Acta 63,

1939–1953.

Wilson, M.J., 1966. The weathering of biotite in some Aberdeenshire

soils. Mineralogical Magazine 35, 1080–1093.

Yoshimura, K., Matsuoka, S., Inokura, Y., Hase, U., 1992. Flow

analysis for trace amounts of copper by ion-exchanger phase

absorptiometry with 4,7-diphenyl-2,9-dimethy1-1,10-phenanthro-

line disulphonate and its application to the study of karst ground-

water storm runoff. Analytica Chimica Acta 268, 225–233.

water percolating behavior indicated by the water

Documents